📊 Data Availability Assessment

May 27, 2026 · View on GitHub

📌 Template Instructions: Copy to analysis/daily/{date}/{article-type}-run{N}/data-availability-assessment.md. This is the FIRST artifact produced in Stage A. Fill every row of the data-source table with the actual tool-call outcome for this run. Set manifest.dataMode based on §4 DataMode Decision. Every Stage-B artifact MUST read this file before deciding which template variant to use.

🎯 Purpose: Standardised triage artifact that records which EP MCP endpoints, IMF API, and other data sources are available for this run. Drives the manifest.dataMode selection that determines whether Stage-B artifacts use full-data templates or degraded-mode variants. Prevents ad-hoc floor reductions and inferred-content disclaimers scattered across individual artifacts.

📋 Document Metadata

Field	Value
Report ID	DAA-YYYY-MM-DD-runNN
Run date	YYYY-MM-DD
Article type	breaking / week-in-review / motions / etc.
Stage	A (pre-analysis)
Produced by	Stage A script / agent first call
Read by	Every Stage-B artifact template selection

1️⃣ Data Availability Status Map

%%{init: {"theme":"dark","themeVariables":{"primaryColor":"#1565C0","primaryTextColor":"#ffffff","primaryBorderColor":"#0A3F7F","lineColor":"#90CAF9","secondaryColor":"#2E7D32","secondaryTextColor":"#ffffff","tertiaryColor":"#FF9800","tertiaryTextColor":"#000000","mainBkg":"#1565C0","secondBkg":"#2E7D32","tertiaryBkg":"#FF9800","noteBkgColor":"#FFC107","noteTextColor":"#000000","errorBkgColor":"#D32F2F","fontFamily":"Inter, Helvetica, Arial, sans-serif"}}}%%
flowchart LR
    START["Stage A<br/>Data Probe"] --> EP_MCP["EP MCP Endpoints"]
    START --> IMF["IMF SDMX 3.0"]
    START --> WB["World Bank MCP"]
    START --> DOCEO["DOCEO XML / Voting"]

    EP_MCP -->|"✅ OK"| EP_OK["Full EP Data"]
    EP_MCP -->|"⚠️ Partial"| EP_PART["Degraded EP"]
    EP_MCP -->|"❌ Fail"| EP_FAIL["EP Unavailable"]

    IMF -->|"✅ live"| IMF_OK["Live WEO"]
    IMF -->|"⚠️ cache"| IMF_CACHE["Stale Cache"]
    IMF -->|"❌ F6"| IMF_FAIL["KB-Estimate Only"]

    DOCEO -->|"✅ < 4 weeks"| RCV_OK["RCV Full"]
    DOCEO -->|"⚠️ lag > 4 weeks"| RCV_LAG["RCV Degraded"]
    DOCEO -->|"❌"| RCV_NONE["RCV Unavailable"]

    EP_OK --> FULL["dataMode: full"]
    IMF_OK --> FULL
    RCV_OK --> FULL

    EP_PART --> DEGRADE["dataMode: degraded-voting<br/>or degraded-imf"]
    IMF_CACHE --> DEGRADE
    RCV_LAG --> DEGRADE

    IMF_FAIL --> FALLBACK["dataMode: degraded-imf"]
    RCV_NONE --> FALLBACK
    EP_FAIL --> MINIMAL["dataMode: minimal"]

    style FULL fill:#2E7D32,color:#ffffff
    style DEGRADE fill:#FF9800,color:#000000
    style FALLBACK fill:#D32F2F,color:#ffffff
    style MINIMAL fill:#7B1FA2,color:#ffffff

2️⃣ Run Configuration Context

Field	Value
EP MCP gateway URL	(from EP_MCP_GATEWAY_URL env var — do not log actual URL; record ✅/❌)
EP MCP server version	1.3.12 (from MCP config)
IMF SDMX probe script	`scripts/imf-mcp-probe.sh` (was it called? ✅/❌)
Pre-fetch script	`scripts/prefetch-ep-feeds.sh` (outcome: ✅ OK / ⚠️ partial / ❌ failed)
Run trigger	workflow_dispatch / schedule / push

3️⃣ Per-Source Triage Table

Required: one row per data source below. Record the actual tool-call outcome for this run, not an expected or default value.

Source	Tool / Endpoint	Called?	Outcome	Records / Size	Lag / Freshness	Impact on Analysis	Mitigation Applied
EP MCP — Voting records	`get_voting_records` (dateFrom / dateTo)	✅/❌	empty / partial / full	N records	Last RCV: YYYY-MM-DD; lag: N weeks	Voting analysis full / degraded / unavailable	Adopted-texts proxy / none
EP MCP — Latest votes (DOCEO)	`get_latest_votes`	✅/❌	empty / partial / full	N records	Last RCV: YYYY-MM-DD; lag: N weeks	As above	As above
EP MCP — Adopted texts	`get_adopted_texts`	✅/❌	available / empty	N texts	Published YYYY-MM-DD	Coalition proxy available / unavailable	Used as voting proxy / none
EP MCP — Procedures feed	`get_procedures_feed`	✅/❌	current / stale (1972–1987) / empty	N items	Staleness label: STALENESS_WARNING?	Procedures data usable / stale proxy	Use adopted-texts / procedures endpoint
EP MCP — Procedures list	`get_procedures`	✅/❌	available / empty	N items	N/A	Supplement to stale feed	Used as proxy for procedures-feed
EP MCP — MEPs	`get_current_meps` / `get_meps`	✅/❌	available / empty	N MEPs	N/A	Seat-count data available / unavailable	None
EP MCP — Plenary sessions	`get_plenary_sessions`	✅/❌	available / empty	N sessions	Latest: YYYY-MM-DD	Session data available / unavailable	None
EP MCP — Committee docs	`get_committee_documents_feed`	✅/❌	available / empty	N docs	Latest: YYYY-MM-DD	Committee analysis available / limited	None
IMF SDMX 3.0 REST	`imf-fetch-data` / `scripts/imf-mcp-probe.sh`	✅/❌	live / F6 / timeout	N datasets	Vintage: WEO-YYYY-MM	Economic context full / fallback	Prior-run cache / KB-estimate
IMF prior-run cache	`data/cache/imf/*.json`	✅/❌	present / absent	N files	Cache date: YYYY-MM-DD	Economic context from cache / none	Uses cache fallback
World Bank MCP	`worldbank-mcp/get-*-data`	✅/❌	available / empty	N records	N/A	Non-economic cross-refs available / none	None

4️⃣ DataMode Decision

Read §3 before selecting dataMode. The selected value is written to manifest.dataMode.

dataMode value	When to use	Floor factor applied by validator
`full`	All primary sources available: EP MCP (including voting records ≤4 weeks old) + IMF API live or cached	1.0
`degraded-voting`	EP MCP available; voting records have publication lag > 4 weeks or are empty	0.85
`degraded-imf`	EP MCP available; IMF SDMX returned F6 / timeout; knowledge-base estimate required	0.85
`title-only`	EP MCP available but most feeds empty; limited analysis possible	0.75
`minimal`	EP MCP gateway unreachable; structural-only analysis	0.65

Selected dataMode for this run: full / degraded-voting / degraded-imf / title-only / minimal

Rationale: ≥30 words explaining why this dataMode was selected based on the triage table above.

Template variants to use in Stage B:

Artifact	Template to use
`intelligence/voting-patterns.md`	`voting-patterns.md` (full) OR `intelligence/voting-patterns.degraded.md` (degraded-voting)
`intelligence/economic-context.md`	`economic-context.md` (full / cache) OR `intelligence/economic-context.fallback.md` (degraded-imf)
`intelligence/mcp-reliability-audit.md`	`mcp-reliability-audit.md` (always full — records availability facts)
All other artifacts	Use standard templates; note any specific source limitations inline

5️⃣ Procedures-Feed Staleness Note

Always populate this section. The get_procedures_feed endpoint consistently returns historical data (1972–1987 era) with a STALENESS_WARNING in dataQualityWarnings. This is a known EP upstream pattern, not a run-specific defect.

Field	Value
`get_procedures_feed` outcome	Stale (1972–1987 tail) / Current / Empty
STALENESS_WARNING present?	Yes / No
Mitigation	Use `get_procedures` (paginated list, current data) + `get_adopted_texts` as procedure-outcome proxy
Impact on analysis	Procedure stage data sourced from `get_procedures` list (reliable) rather than feed (stale)
Cross-reference	See `intelligence/procedures-proxy.md` for full staleness-mitigation protocol

6️⃣ IMF API Status

Field	Value
IMF SDMX probe status	live / F6 / timeout / not-attempted
Probe script run?	✅ `scripts/imf-mcp-probe.sh` ran / ❌ not run (explain why)
Cache available?	Yes — `data/cache/imf/[file].json` (cache date YYYY-MM-DD) / No
Fallback template required?	Yes → use `intelligence/economic-context.fallback.md` / No → use `economic-context.md`

7️⃣ Voting Data Freshness Summary

Field	Value
Latest RCV date found	YYYY-MM-DD OR "none in window"
Lag to run date	N weeks
Freshness threshold	4 weeks (per osint-tradecraft-standards.md §3.1)
Degraded-voting triggered?	Yes (lag > 4 weeks or empty) / No
Degraded template required?	Yes → use `intelligence/voting-patterns.degraded.md` / No → use `voting-patterns.md`

8️⃣ Cross-Run Continuity

Field	Value
Prior run ID	Previous run ID or "none"
Prior run dataMode	full / degraded-voting / degraded-imf / etc.
Prior run IMF cache available?	Yes / No
Carry-forward data	List any prior-run cache files used this run (cite path + vintage)

9️⃣ Quality Assessment

Source	Quality	Rationale
EP MCP endpoints	🟢/🟡/🔴	State how many endpoints returned non-empty results
Voting records	🟢/🟡/🔴	Fresh (≤4 weeks) / lagged (4–8 weeks) / unavailable
IMF data	🟢/🟡/🔴	Live / stale cache / KB-estimate only
Procedures data	🟢/🟡/🔴	Current (via `get_procedures`) / stale feed only
World Bank data	🟢/🟡/🔴	Available / partial / unavailable

Overall run quality: 🟢 Full / 🟡 Degraded / 🔴 Minimal

🔟 Manifest Update Required

After writing this artifact, add these fields to manifest.json:

{
  "dataMode": "<selected-dataMode>",
  "dataAvailabilityAssessment": "data-availability-assessment.md",
  "votingDataFreshness": {
    "latestRcvDate": "YYYY-MM-DD",
    "lagWeeks": 0,
    "degraded": false
  },
  "imfApiStatus": {
    "status": "live",
    "cacheAvailable": false,
    "fallbackRequired": false
  }
}

🔗 Cross-References

../methodologies/reference-quality-thresholds.json — degradedFloorFactor per dataMode (read by scripts/validate-analysis-completeness.js)
intelligence/voting-patterns.degraded.md — use when degraded-voting
intelligence/economic-context.fallback.md — use when degraded-imf
intelligence/procedures-proxy.md — use when procedures feed is stale
../methodologies/osint-tradecraft-standards.md §3.1 — degraded-source protocol
.github/prompts/07-mcp-reference.md §11 — Audit-Recurring Items Triage Table

Document Control: analysis/templates/data-availability-assessment.md · Template v1.0 · Stage: A · Required for every article-generating run · Read by every Stage-B artifact for template-variant selection.